🛡️ Sentinel - anti-scam/spam defense for Discord

Scammers repeat themselves. Sentinel notices.

A bot joins, pastes the same free-Nitro link into every channel it can see, and by the time someone with a ban button is awake, a few of your members have already clicked it.

That repetition is the one thing a scam bot can't avoid, and it's what Sentinel watches for. No leveling, no music, no dashboard to babysit. It does anti-scam moderation and nothing else.

How it catches things

Repeated spam. The same message across several channels within a few seconds, or the classic @everyone + screenshot blast. Default is 3 messages in 30 seconds spanning at least 2 channels, and both numbers are yours to change. It matches on the text itself, so link scams with no image get caught the same way.

Known phishing domains. Anything on the FishFish blocklist is removed on the first message. No repetition needed, no waiting. Edited messages get re-checked, so nobody slips a link in after the fact. Add your own blocked domains, or allow one that FishFish flags and you actually trust.

Honeypot channels. Mark a channel as auto-ban. Scam bots post in every channel they can see, so they tend to walk straight into it and ban themselves.

It won't ban your moderators

Bots, and anyone with Administrator, Manage Messages or Ban Members, are exempt from automatic bans. That isn't a setting you have to remember to turn on.

If a mod account does start spamming (such as from the account being compromised), Sentinel deletes the messages and alerts you instead of banning a victim. It can time the account out too, if you'd rather.

When it does get one wrong, every ban alert in your mod log carries an Unban button. One click.

Not ready to let it act at all? /antispam action observe logs every would-be action and touches nobody. Run it for a week, see what it would have done, then decide.

Privacy

Message content is never logged or written to disk. Detection keeps short-lived hashes in memory so it can tell a message is being repeated, and that's the whole of it. Sentinel doesn't ask for audit log access, and it never asks for Administrator.

Commands

• /help - what Sentinel does and how to set it up
• /antispam - thresholds, what happens when one trips, the phishing filter, your domain lists, and where logs go
• /autoban - add, remove and test honeypot channels
• /sentinel check - tells you whether anything is stopping Sentinel doing its job
• /stats - what it's caught in your server
• /safety - scam-avoidance resources you can post for your members
• Report scam - right-click any message to flag it to your mod log

Getting started

Invite it. Detection is on from the moment it joins and the defaults are fine for most servers, so you can genuinely stop there. If you do want to tune it, everything is configured per server with slash commands.

🔗 Website & docs: https://sentinelbot.app 📄 Privacy policy: https://sentinelbot.app/privacy ☕ Support development: https://buymeacoffee.com/thesentinelbot

Free, in as many servers as you like.